Glucagon-like-peptide (GLP-1) can delay gastric emptying and increase the risk of aspiration during esophagogastroduodenoscopy (EGD). However, there is no consensus guidelines on whether to discontinue GLP-1 prior to EGD. We aim to assess rates of repeat EGD in non-diabetic adults with GLP-1 prescription.
Method
This is a retrospective cohort study performed using a de-identified database, TriNetX (Cambridge, MA) which gathers data from several healthcare systems across the United States. Individuals with body mass index (BMI) of ≥ 27 kg/m2 without diabetes mellitus who underwent diagnostic EGD with/ without brushing or biopsy (CPT code 43235, 43239) since 2014 at least 1 week after GLP -1 (semaglutide or liraglutide) prescription were included as the study group. A cut-off ≥ 27 kg/m2 was selected as GLP-1 are indicated for weight management in patients with BMI ≥ 27 kg/m2. EGDs were reviewed from 2014 as GLP-1 was approved for weight loss in 2014. Individuals with obesity without diabetes mellitus who were never prescribed GLP-1 were considered as control group. Individuals who underwent therapeutic endoscopic interventions were excluded. Similarly diagnoses that predispose to mechanical or functional obstruction including malignancy of esophagus, stomach, duodenum, pancreas, achalasia were also excluded.
Primary aim was to compare repeat endoscopic evaluation between two groups. This was identified as repeat EGD within 60 days. Secondary aim was to compare rates of gastric emptying study and diagnosis of gastroparesis (surrogates for food retained in stomach) within a year of EGD.
Propensity matching with age, gender, BMI, use of opioid analgesics, anti-depressants, octreotide, diagnoses of esophagitis, peptic ulcer, gastric ulcer were performed between the two cohorts. Cohorts were compared using odds ratio calculation with TriNetX Advanced Analytics.
Results
Out of 16,995,005 individuals with a recorded BMI, 8,767,194 (51.6%) individuals had BMI ≥ 27 kg/m2 among which 2,200 were prescribed GLP-1 and then underwent ≥ 1 EGDs since 2014. Control group comprised of 260,090 adults who were never prescribed GLP-1. Cohorts were propensitymatched in Table 1. In the study group, 48 (2.2%) had repeat EGD within 60 days compared to 6,059 (2.3%) in the control group. The odds ratio of matched cohorts was 0.80 (0.54, 1.17). 38 adults (1.7%) with GLP-1 underwent GES compared to 3,951 (1.4%) with an odds ratio of matched cohorts as 0.66 (0.43, 1.0). Similarly, 22 adults (1%) in study group had a subsequent diagnosis of gastroparesis compared to 2,450 (0.9%) in control group resulting in an odds ratio of 0.78 (0.44, 1.38) (Table 2).
Conclusion
Using a population-level database, we did not identify any significant difference in the rates of repeat upper endoscopy or diagnosis of gastroparesis after an initial endoscopy in non-diabetic adults who were prescribed GLP-1.

